Multi-scale turbulence-chemistry closure in large-eddy simulation to account for sensitivity to fuel composition and properties fuel composition and properties manifest themselves in the fuel…

This is a $1,196,288.99 cooperative agreement awarded by the Shared Services Center, a civilian agency, to the Georgia Tech Research Corporation (Georgia Tech) to develop a new high-fidelity modeling approach for predicting fuel sensitivity and pollutant emissions in turbulent combustion systems. The effort will integrate a hybrid spray simulation method, a subgrid turbulent mixing model, and reduced kinetics/transport models to account for the effects of fuel composition and properties on phenomena like lean blowout, cold start, and particulate emissions. This work aims to enhance the Large Eddy Simulation (LES) capabilities for turbulent reacting flows. There is no set-aside designation for this contract. Georgia Tech, a non-profit research organization and state institution of higher education, has extensive experience delivering R&D, engineering, and technical services to various federal agencies through prior contracts and grants.

Multi-scale turbulence-chemistry closure in large-eddy simulation to account for sensitivity to fuel composition and properties fuel composition and properties manifest themselves in the fuel properties, reaction kinetics and in the resulting species thermodynamic/transport properties. Additionally, multi-component liquid fuel with widely different properties such as viscosity, boiling points and can also have different spray patterns from the same injector. Predicting fuel sensitivity then becomes challenging since validated kinetics/transport for different types of fuels have to be available for modeling and experimental evaluation is needed to show this effect under turbulent combustion conditions of practical interest. In particular, fuel sensitivity of the prediction of lean blowout (LBO), cold start, altitude-relight and particulate emissions under operating conditions of relevance is needed and is the focus of this modeling study using a state-of-art large-eddy simulation (LES) method. However, conventional LES only captures features larger than the grid and thus, the spray formation and all of the chemical sensitivities and transport effects that occur at the small scales require modeling. The proposed effort will bridge these scales by explicitly combining a new hybrid eulerian-eulerian (EE) and eulerian-lagrangian (EL) spray simulation method developed recently at georgia tech with a subgrid turbulent mixing approach based on the linear eddy mixing (LEM) model. Spray breakup models to be developed under the national jet fuels combustion program (njfcp) and/or NASA TTT 2.1.8 project will be integrated within the LES solver. Furthermore, within the subgrid a new eulerian two-phase LEM closure will be implemented and validated for near injector flow modeling. Finally, the reduced/skeletal kinetics and multi-species transport properties to be obtained in topic 2.1.9 (and/or the njfcp program) will be implemented to explicitly include finite-rate kinetics for different fuels. Kinetics for both NOX and soot will be included and especially for soot modeling a subgrid method of moment with interpolative closure (momic) will be used in LEM to circumvent some of the known limitations of soot modeling in conventional LES closures. This combined multi-scale approach will create a new high fidelity model capable of predicting the fuel sensitivity and pollutant emission in actual combustion systems. Validation for chosen fuels, such as the CAT A2, C1 and C5 (or other fuels with available kinetics) will be carried out using canonical flame data and simulation in the actual njfcp area 6 (univ. Dayton udri) and the NASA 9-point lean-direct injection (LDI) rigs will be carried out for final evaluation and assessment. Although most of the basic tools are operational, the integration of these various closures is needed to predict sensitivity of LBO to fuel properties and to model the initiation of ignition kernel and its propagation and ignition of cold liquid fuel under the other conditions of interest. Performance acceleration is also needed and will be achieved by a combination of preconditioning and hybrid MPI-openmp programing (being developed in other projects). Additional speedup of kinetics and multi-species transport will be accomplished in collaboration with researcher in topic 2.1.9 and njfcp to integrate adaptive kinetics, lumped transport models and stiffness removal approaches. Finally, all the subgrid models will be created as user defined function (UDF) modules so that they can be delivered to NASA and to industry. The LES code and all the data (in an archival site) will also be available to NASA if requested (consistent with past effort where we delivered our codes in earlier projects).